Safe Flight Envelope Characterization and Emergency Landing with Reachability
Abstract
The purpose of this action is to provide funds for a new start Grant. FY23 funding in the amount of $54,504.45 and FY24 CR3 funds in the amount of $20K is provided.--We propose to develop methods to efficiently compute unsafe operating flight conditions for which no control sequence exists to safely land the helicopter in the event of partial or total engine failure. A chart called a height-velocity diagram or H-V diagram needs to be produced for each airframe, and a human pilot can utilize the diagram to avoid operating in unsafe conditions. If operatingin the safe region, there exists a control sequence to initiate an autorotation, whereby the helicopter enters a glide slope so as to maintain rotor inertia and then flare to slow down near the ground and land safely. H-V diagrams are ultimately determined through flight testing, which is inherently dangerous since the test objective is to define the unsafe boundaries of flight operations. Therefore, there is a need to accurately compute the safe and unsafe regions directly from the helicopter dynamics, and construct a H-V diagram prior to flight testing. The capability to produce H-V diagrams isdesired for aircraft that are in the early design stagesand have not yet flown, aircraft that are already flying and preparing for an H-V flight test event, as well as refining the H-V diagram of operational aircraft. The capability is required for aircraft that do not have a representative, high-fidelity flight simulation, as well as for more mature aircraft for which a high-fidelity simulation is available.We consider the safe region as any initial flight condition where there exists a control sequence that can steer the system to a safe landing condition, i.e. minimal vertical and horizontal velocity at the ground level, rotor near level, etc. Determining the set of states of a dynamical system that canbe driven into a particular final condition is commonly referred to as reachability analysis, and reachable sets can be determined from the sub-zero level set of the viscosity solution to a Hamilton#Jacobi (HJ) partial differential equation (PDE). Traditionally, these HJ PDEs are solved numerically by constructing a dense discrete grid of the solution space, and are supported by mature theory. Despite this, HJ reachability analysis has suffered one critical draw-back: Computing the elements of a spatial grid scales poorlywith dimension, and therefore have limited applicability for vehicle problems where the dimension of the state space is greater than four.We propose to avoid a spatial grid of the state space and instead form a grid over time, where numerical solutions are obtained by constructing a trajectory optimization problem the coincides with viscosity solution of the reachability HJ PDE. We consider the case where the time-to-land need not be known a priori by considering that the intersection of the helicopter with the ground could occur at any time on the interval zero to infinity. From this we compute the optimal landing sequence to autonomously achieve a safe landing from any initial condition that is within the backwards reachable tube.The method will be tested and verified against the Manned Flight Simulator (MFS), developed by the Flight Vehicle Modeling and Simulation branch at the Naval Air Warfare Center Aircraft Division (NAWCAD). This also includes the collection of human response data from trained pilots as ensure that any safe landingmaneuver is feasible for a human pilot to execute. An additional goal of the project is to develop a human-in-the-loop pilot queuing tool that can guide a pilot to a safe landing on a feasible trajectory during an emergency landing. This tool is then to be testedin the MFS. Additionally, small scale hardware implementation is proposed to test and verify the methods. As part of the effort, failure detection techniques will be investigated that could robustly detect an engine failure.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- May 15, 2024
- Source ID
- N000142412322
Entities
People
- Matthew Kirchner
Organizations
- Auburn University
- Office of Naval Research
- United States Navy